I'm going to do a simonw here and coin "Markdown hoarding" for the Claudeism of producing reams of hyper-dense prose and compounding it with every commit that touches docs. The documentation gets more and more bloated.

The default Claude voice is breathless, jargon-heavy showboating:

    WebRTC transport (trickle ICE) — RTCPeerConnection data channels with trickle signaling; same-machine mDNS-loopback rewrite. Browser/device-verified. peerd-distributed/transport/transports/webrtc.js

    TOFU rootfs integrity pin — pure decision logic: trust-on-first-use pin of base-image size + first-64KiB SHA-256, verified each boot, fails closed on drift. Head-only by construction — a faithful-head/tampered-tail host is an acknowledged residual gap. peerd-engine/image-pin.js (partial)

    PDF reading (read_pdf) — pdf.js text-layer extraction with opt-in SRI-pinned OCR engine + page assembly. pdf/, tools/defs/read-pdf.js

Agree on markdown bloat/hoarding. Appreciate the feedback, already doing a pass based on other feedback and trimming fat.